A Unique Presentation of Infantile-Onset Colitis and Eosinophilic - - PDF document

ORIGINAL ARTICLE

A Unique Presentation of Infantile-Onset Colitis and Eosinophilic Disease without Recurrent Infections Resulting from a Novel Homozygous CARMIL2 Variant

Alina Kurolap1,2 & Orly Eshach Adiv2,3 & Liza Konnikova4,5,6 & Lael Werner7,8 & Claudia Gonzaga-Jauregui9 & Maya Steinberg1 & Vanessa Mitsialis5,6 & Adi Mory1 & Moran Y. Nunberg7,8 & Sarah Wall5 & Ron Shaoul2,3 & John D. Overton9 & Regeneron Genetics Center & Alan R. Shuldiner9 & Yaniv Zohar2,10 & Tamar Paperna1 & Scott B. Snapper5,6,11 & Dror S. Shouval7,8 & Hagit Baris Feldman1,2

Received: 16 September 2018 /Accepted: 14 April 2019 /Published online: 11 May 2019 # Springer Science+Business Media, LLC, part of Springer Nature 2019

Abstract Purpose This study aimed to characterize the clinical phenotype, genetic basis, and consequent immunological phenotype of a boy with severe infantile-onset colitis and eosinophilic gastrointestinal disease, and no evidence of recurrent or severe infections. Methods Trio whole-exome sequencing (WES) was utilized for pathogenic variant discovery. Western blot (WB) and immuno- histochemical (IHC) staining were used for protein expression analyses. Immunological workup included in vitro T cell studies, flow cytometry, and CyTOF analysis. Results WES revealed a homozygous variant in the capping protein regulator and myosin 1 linker 2 (CARMIL2) gene: c.1590C>A; p.Asn530Lys which co-segregated with the disease in the nuclear family. WB and IHC analyses demonstrated reduced protein levels in patient’s cells compared with controls. Moreover, comprehensive immunological workup revealed severely diminished blood-borne regulatory T cell (Treg) frequency and impaired in vitro CD4+ T cell proliferation and Treg

• generation. CyTOF analysis showed significant shifts in the patient’s innate and adaptive immune cells compared with healthy

controls and ulcerative colitis patients. Conclusions Pathogenic variants in CARMIL2 have been implicated in an immunodeficiency syndrome characterized by recur- rent infections, occasionally with concurrent chronic diarrhea. We show that CARMIL2-immunodeficiency is associated with significant alterations in the landscape of immune populations in a patient with prominent gastrointestinal disease. This case provides evidence that CARMIL2 should be a candidate gene when diagnosing children with very early onset inflammatory and eosinophilic gastrointestinal disorders, even when signs of immunodeficiency are not observed.

Alina Kurolap and Orly Eshach Adiv contributed equally to this work. Dror S. Shouval and Hagit Baris Feldman contributed equally to this work. Electronic supplementary material The online version of this article (https://doi.org/10.1007/s10875-019-00631-6) contains supplementary material, which is available to authorized users. Journal of Clinical Immunology (2019) 39:430–439 https://doi.org/10.1007/s10875-019-00631-6 * Hagit Baris Feldman hb_feldman@rambam.health.gov.il

1

The Genetics Institute, Rambam Health Care Campus, Haifa, Israel

2

The Ruth and Bruce Rappaport Faculty of Medicine, Technion – Israel Institute of Technology, Haifa, Israel

3

Pediatric Gastroenterology, Rambam Health Care Campus, Haifa, Israel

4

Devision of Newborn Medicine, Department of Pediatrics, Children’s Hospital of Pittsburgh of UPMC, Pittsburgh, PA, USA

5

Division of Gastroenterology, Hepatology and Nutrition, Boston Children’s Hospital, Boston, MA, USA

6

Harvard Medical School, Boston, MA, USA

7

Pediatric Gastroenterology Unit, Edmond and Lily Safra Children’s Hospital, Sheba Medical Center, Tel Hashomer, Israel

8

Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel

9

Regeneron Genetics Center, Tarrytown, NY, USA

10

Institute of Pathology, Rambam Health Care Campus, Haifa, Israel

11

Division of Gastroenterology, Hepatology and Endoscopy, Brigham and Women’s Hospital, Boston, MA, USA

Keywords CARMIL2 . RLTPR . infantile colitis . very early onset inflammatory bowel disease . immunodeficiency

Introduction

Infantile-onset inflammatory bowel disease (IBD) refers to a rare group of inflammatory gastrointestinal disorders with onset be- fore the age of 2 years [1, 2]. To date, nearly 100 different monogenic conditions have been described that include gastro- intestinal manifestations with an IBD-like phenotype among

• ther systemic features [3]. Considering the intestines are a ma-

jor lymphoid organ, primary immunodeficiency and immune dysregulation syndromes contribute to more than half of mono- genic IBD syndromes. Suspicion for immune-based hereditary IBD disorder is raised by a clinical history of recurrent infec- tions, young age at presentation, severe manifestations, consan- guinity, and immunological workup impairments [1, 3]. Recently, several patients with primary immunodeficiency have been described with pathogenic variants in the capping protein regulator and myosin 1 linker 2 (CARMIL2), also known as RGD, leucine-rich repeat, tropomodulin, and proline-rich-containing protein (RLTPR) [4–7]. CARMIL2- immunodeficiency syndrome is characterized by a combined immune defect in Tand B cells, as shown by various studies in humans and mice [8]. The limited number of patients reported presented with recurrent infections, mostly of the respiratory system, and cutaneous features, including psoriatic rash, ecze- ma, and skin warts. Chronic diarrhea was observed in some of the patients, though there is limited related phenotypic data [4–7]. We report on a patient with a novel homozygous CARMIL2 variant, manifesting primarily as infantile-onset co- litis and eosinophilic disease.

Materials and Methods

Study Participants

The study was approved by the institutional Helsinki commit- tee, and written informed consent was obtained as customary. The studied family included an affected child, both his par- ents, and a healthy brother.

Genetic Analysis

BTrio^ whole-exome sequencing (WES) was performed in collaboration with the Regeneron Genetics Center (RGC). Protein-coding regions were captured using the IDT xGen capture platform (Integrated DNA Technologies, Coralville, IA, USA) and sequenced on the Illumina HiSeq2500 platform (Illumina, San Diego, CA, USA). We utilized the Genoox data analysis platform Ltd. (Tel Aviv, Israel) and the RGC bioinformatics pipeline for mapping and alignment of the ob- tained sequence reads to the human genome reference assem- bly, variant calling and annotation, and subsequent data analysis. We filtered WES data for rare (defined as variants with a minor allele frequency [MAF] < 0.01 in unaffected controls from population databases, such as gnomAD [9], dbSNP [10], 1000Genomes [11], Greater Middle-East Variome [12], the Rambam Genetics Institute in-house database of 1000 Israeli exomes, and the internal RGC database) coding, protein- altering variants (missense, nonsense, frameshift, and splice site). We investigated all possible disease-causing variants identified by WES, including de novo mutations, X-linked, and compound heterozygous and homozygous variants under a recessive mode of inheritance due to consanguinity between the parents. Pathogenicity of the mutations was assessed by various in silico programs, including SIFT, MutationTaster, and PolyPhen-2 [13–15]. The ConSurf server and SWISS- MODEL were used for protein modeling and variant assess- ment [16, 17]. The candidate variant in CARMIL2 was validated and test- ed for segregation with disease in the healthy brother by Sanger sequencing on an ABI Prism 3500 Genetic Analyzer (Applied Biosystems, Waltham, MA, USA), using 5′-AGAC CACACATTGGGAGAGG-3′ forward and 5′-ACCG GACGTTGAAGTTCCTT-3′ reverse primers.

RNA Analysis

RNA was extracted from peripheral blood of the patient and a healthy control, using the PureLink RNA Mini kit (Invitrogen, Carlsbad, CA, USA) and reverse transcribed with the High- Capacity cDNA Reverse Transcription Kit (Applied Biosystems), following the manufacturer’s protocols. Patient and control cDNA samples and a wild-type gDNA control were analyzed by standard PCR using primers designed to flank over exons 14–21 (forward 5′-CTGAGCCGTCCTAA CGTACT-3′, reverse 5′-ACCCAAAGCAGATGTGTGGT-3′)

• f CARMIL2. The expected cDNA and gDNA sizes were 765

and 1779 bp, respectively. PCR products were subjected to 4% agarose gel electrophoresis with HyperLadder 100 bp (Bioline, London, UK) and Sanger-sequenced to observe the candidate variant effect on splicing.

Western Blotting

Total protein was extracted from patient and healthy control lymphoblastoid cell lines (LCLs) using RIPA buffer and sub- jected to SDS-PAGE using standard protocols. The EM-53

J Clin Immunol (2019) 39:430–439 431

anti-CARMIL2 antibody (#11–718, ExBio, Vestec, Czech Republic; 1:500 dilution) was used for protein-of- interest detection, and anti-α-Tubulin (#T9026, Sigma- Aldrich, St. Louis, MO, USA; 1:5000 dilution) was used as a loading control. Results were visualized on ImageQuant LAS 4010 (GE Healthcare, Waukesha, WI, USA) and ana- lyzed using NIH ImageJ software.

Immunohistochemistry

Formalin-fixed paraffin-embedded 4-μm sections from pa- tient and age- and gender-matched control colonic and esoph- ageal samples were deparaffinized and stained using the EM- 53 anti-CARMIL2 antibody (ExBio) diluted at 1:100. Staining was performed on the Ventana Benchmark ULTRA system (Roche Diagnostics, Risch-Rotkreuz, Switzerland). Normal tonsil biopsy was used as a positive control.

Isolation of Peripheral Blood Mononuclear Cells

Blood was collected in EDTA tubes. Peripheral blood mono- nuclear cells (PBMCs) were isolated by Ficoll-Paque PLUS (GE Healthcare, Little Chalfont, UK) gradient, according to the manufacturer’s instructions.

In vitro T cell Proliferation

CD4+CD45RA+CD25− naïve T cells were isolated by magnetic-activated cell sorting (MACS) beads (Miltenyi Biotec, Bergisch Gladbach, Germany). Isolation purity was more than 95%. To assess proliferation, we cultured 5 × 105 mM carboxyfluorescein succinimidyl ester (CFSE)–la- beled naïve T cells with soluble αCD3/αCD28-coated beads (ThermoFisher Scientific, Waltham, MA, USA) and IL-2 (100 ng/mL, Peprotech, Rocky Hill, NJ, USA). Proliferation was assessed after 4 days based on CFSE dilution and deter- mined by flow cytometry.

In vitro Treg Generation

MACS-sorted CD4+ naïve T cells were stimulated with αCD3/αCD28 beads, TGFβ 2 ng/mL (Peprotech), and IL-2 (100 ng/mL) for 4 days. Cells were then fixed, permeabilized, and stained for CD25 and FOXP3.

Flow Cytometry

Cells were acquired using Navios flow cytometer (Beckman- Coulter, Brea, CA, USA) and analyzed using FlowJo. For FOXP3 staining, fixation/permeabilization kit by eBioscience was used. The following antibodies were used: CD4 clone RPA-T4 (BD Bioscience), CD25 clone 2A3 (BD Bioscience), and FOXP3 clone PCH101 (eBioscience). All flow cytometry experiments were performed in duplicate and a representative image was chosen.

CyTOF Studies

PBMCs from the CARMIL2-immunodeficiency patient, six healthy controls (ages 5–18) and two patients with active ul- cerative colitis (UC, ages 18–22), were either stained directly as subsequently described or stimulated for 4 h with 50 ng/mL phorbol myristate acetate (PMA) and 500 ng/mL ionomycin and then stained with a panel of metal-tagged antibodies targeting markers of major immune cell lineages and various

• cytokines. The samples were run on Helios2 mass cytometer

(Fluidigm, San Francisco, CA, USA). FCS files obtained were analyzed with premium Cytobank software and pre-gated on CD45+/viable/single/DNA+ events before initiating the anal-

• ysis. Normalization beads were used and gated out of the
• analysis. The data was automatically clustered with

Rphenograph and cluster identity was manually labeled based

• n marker expressed in the individual clusters. The data were

also independently reduced on ViSNE plots via Cytobank,

• nce again with manual labeling of cellular populations based
• n marker expression. Cluster percentage (% of CD45+ viable

single events) was computed and plotted for comparison be- tween patient groups using Prism7 software. Cytokine levels were analyzed either by median intensity of the marker within particular clusters or by percent-positive events within that cluster.

Results

Case Description

A boy was referred for genetic evaluation at the age of 10 years due to infantile-onset colitis. He was born to healthy parents of Christian Arab descent, who are second-degree cousins (Fig. 1a). The patient presented at age 6 months with chronic bloody diarrhea, accompanied by recurrent fever, aphthous stomatitis, psoriatic rash, severe failure to thrive (FTT), mal- nutrition, and subsequent physical and motor developmental

• delays. He did not have recurrent or severe infections that

would suggest a primary immunodeficiency. Colonoscopy at the age of 4 years revealed severe active chronic colitis with skip lesions, and hematoxylin and eosin staining demonstrated massive eosinophilic infiltration of the lamina propria of healthy and inflamed colonic mucosa (Fig. 1b). Skin biopsy was consistent with a diagnosis of psoriasis. He was treated with systemic and topical steroids with partial response, and due to FTT and ongoing symptoms, total parenteral nutrition was given for several months; no other medications were used in view of parental objection. His psoriatic rash resolved at 5 years and has not recurred since.

432 J Clin Immunol (2019) 39:430–439

At 5 years of age, he underwent a sub-total colectomy due to perforation that occurred during routine colonoscopy, leaving him with a protective ileostomy and a rectal pouch. Since then, he suffers from continuous rectal discharge, with repeated rectal biopsies showing eosinophilic proctitis. Over the years, he also developed eosinophilic esophagitis. Amino acid–based formula nutrition combined with topical steroid treatment (oral gel and rectal suppository) was ini- tiated with relatively good results, both clinically and in follow-up esophageal histology. Currently, the boy is thriv- ing properly in terms of height (6th percentile), weight (27th percentile), and cognition. He does not suffer from recurrent infections or exhibits cutaneous findings. His main medical issues revolve around his stoma, rectal dis- charge, and the fact that he is dependent on a hypoaller- genic formula. Eosinophilic infiltration and inflammation are still observed on repeat biopsies.

Identification of a Novel CARMIL2 Candidate Variant

No pathogenic variants were observed in IBD-related genes, such as XIAP, FOXP3, and IL10RA/B [1]. BTrio^ WES anal- ysis identified nine homozygous, three compound heterozy- gous, four X-linked, and no de novo rare protein-altering var- iants that passed the filtering criteria described above (Supplementary Table S1). Of these, we prioritized one can- didate homozygous novel missense variant, which we consid- ered likely related to the patient’s phenotype. The variant oc- curs in exon 17 of CARMIL2 (NM_001013838.1): c.1590C>A; p.Asn530Lys (Fig. 1a,c); the amino acid substi- tution affects the 11th leucine-rich repeat (LRR) domain of the protein (Fig. 1d). The variant has not been reported in general population variant databases or in our internal databases, and it is predicted deleterious by all pathogenicity bioinformatic algorithms used. The variant is located 2-bp upstream of

• Fig. 1 A missense variant in CARMIL2 causes infantile colitis. a Pedigree

and genotypes of studied family. The c.1590C>A variant segregates as expected for an autosomal recessive disease. b Histopathological section

• f patient’s colon tissue, stained with hematoxylin and eosin, revealed

skip lesions of healthy and inflamed tissue. Left panel depicts × 40 magnification of inflamed colonic mucosa showing marked increase in eosinophils (in red) in the lamina propria (up to 100 eosinophils per high- power field). Right panel depicts × 40 magnification of healthy colonic mucosa showing granulation tissue with multiple eosinophils (in red). Zoom-in insets are provided for each section to highlight the eosinophilia. c Schematic representation of the CARMIL2 gene (top) and protein (bottom), highlighting the missense variant described in this report (c.1590C>A; p.Asn530Lys) and previously described variations in the

• gene. The c.1590C>A variant affects the second to the last nucleotide of

exon 17, causing an asparagine-to-lysine substitution at position 530 of the protein. PH, pleckstrin homology; LRR, leucine-rich repeat; HD, helical dimerization; CBR, capping protein-binding; PRD, proline-rich

• domain. d 3D homology model of the CARMIL2 LRR region, composed
• f 16 leucine-rich repeats (modeled by SWISS-MODEL). Asn530 is

located within a highly conserved region in LRR 11. It is predicted to be exposed and functional (source: ConSurf server). b, buried; e, exposed; f, functional, s, structural J Clin Immunol (2019) 39:430–439 433

exon’s 17 donor splice site and hence we considered possible that it could also affect mRNA splicing (Fig. 1c). However, PCR analysis and Sanger sequencing of the proband’s cDNA flanking CARMIL2 exons 14–21 revealed the expected mis- sense variant of c.1590C>A, without any effect on mRNA splicing (results not shown).

CARMIL2 Protein Expression

Western blot analysis revealed a marked reduction of 73% in CARMIL2 protein in patient LCLs compared with a healthy control (Fig. 2a). Immunohistochemical staining for CARMIL2 expression in normal tonsillar tissue (used as positive control) revealed intense staining in endothelial cells, moderate staining in parafollicular lymphocytes, and weak staining in germinal centers (Fig. 2b). The colonic control sample showed positive cytoplasmic staining in lymphocytes, lamina propria plasma cells, crypt ep- ithelium, and endothelium (Fig. 2c), whereas patient samples revealed a lack of staining in lymphocytes and weak, focally positive staining in colonic epithelium and plasma cells (Fig. 2d). No staining was detected in control esophageal squamous epithelium, while intense staining was observed in esophageal immune infiltrates (Supplementary Fig. S1). The patient esoph- ageal sample showed weak staining only in endothelial cells in granulation tissue (Supplementary Fig. S1), compared with in- tense staining observed in control endothelial cells (Fig. 2b).

Immunological Analyses

Blood tests at the age of 12 years demonstrated normal immu- noglobulin levels. The lymphocyte sub-population assay was

• verall normal with slightly elevated B cells and decreased

CD4+ T cells (Table 1). The level of T cell receptor excision

• Fig. 2 CARMIL2 protein expression analyses. a Western blot analysis of

CARMIL2 reveals 27% protein expression in patient LCLs compared with control LCLs. The figure shows a representative blot from n = 2 independent experiments, each performed in duplicate. b Normal tonsillar tissue was used as positive control for CARMIL2 immunohistochemical staining, showing a rage of staining intensities in different cell types: intense staining in endothelial cells, moderate staining in parafollicular lymphocytes, and weak staining in germinal centers. c Colonic sample from age- and gender-matched control shows positive cytoplasmic staining in lymphocytes (*), lamina propria plasma cells (¥), and crypt epithelium (¤). d Patient’s colonic sample shows lack of staining in lymphocytes (*) and weak focally positive staining in colonic epithelium (¤) and plasma cells (¥). All immunohistochemical images are depicted at × 20 magnification 434 J Clin Immunol (2019) 39:430–439

circles, a marker of Tcell maturation, was normal, as was Tcell receptor Vβ repertoire. Previous studies suggested abnormal T cell proliferation among patients with CARMIL2-immunodefi-

• ciency. T cell proliferation in response to phytohemagglutinin

(PHA) was comparable between patient and control, but re- sponse to anti-CD3 was reduced in the patient (Table 1), in contrast to previous reports [4, 6]. In more detailed analyses, naïve CD4+ T cells of the patient exhibited diminished prolif- eration in response to anti-CD3/CD28 and IL-2 stimulation (Fig. 3a). Moreover, in vitro regulatory CD25highFOXP3+ T cell (Tregs) generation was impaired in the patient cells com- pared with control (Fig. 3b). A marked decrease in Treg frequen- cy was also observed in peripheral blood of the CARMIL2- immunodeficiency patient (Fig. 3c). Overall, our studies are in line with previous findings and suggest a marked impairment in T cell activation in the presence of a mutated CARMIL2.

CyTOF Analysis of PBMCs

To obtain a more comprehensive overview of the effect of ab- errant CARMIL2 on different immune populations in the blood, we performed CyTOF analysis of PBMCs obtained from the patient and compared the profiles to healthy controls and pa- tients with UC (Fig. 4, Supplementary Fig. S2). There were significant shifts in both innate and adaptive immune cells in the patient compared with controls and UC patients. Our analyses demonstrate that the patient has increased the abun- dance of effector memory and naïve CD8+ T cells, an increase in naïve CD4+ T cells and transitional B cells (CD38+/CD24+) compared with controls and patients with UC. In contrast, a significant reduction of Tregs, memory B cells (CD27+), NK cells, and monocytes was observed in the patient’s blood (Fig. 4, Supplementary Fig. S2). We next evaluated cytokine production in response to stimulation with PMA and

• ionomycin. Overall, cytokine levels were comparable between

the groups, including IL-21, IL-22, IL-17, TNF-α, and IL1β (data not shown). However, interferon-γ (IFN-γ) production in the CARMIL2-immunodeficiency patient was increased com- pared with controls in several immune subsets, particularly ac- tivated CD8+ T cells (Fig. 5).

Discussion

We describe a patient with infantile-onset colitis, who was found to harbor a homozygous missense variant in CARMIL2 (p.Asn530Lys). The previously described immu- nodeficiency syndrome caused by CARMIL2 mutations in 11 unrelated families consists predominantly of recurrent in- fections and cutaneous features; only six of the 21 published patients were reported to have chronic diarrhea, among the

• ther symptoms, and only one patient was described with

Table 1 Immunological workup

• f the CARMIL2-

immunodeficiency patient Laboratory Patient’s value Normal range Immunoglobulins IgG 806 mg/dL 720–1670 mg/dL IgA 292 mg/dL 63–304 mg/dL IgM 180 mg/dL 49–183 mg/dL Lymphocyte subsets WBC 10,330 cells/m3 Lymphocytes 35.8% CD3+ T cells 64% 60–85% CD4+ T cells 26% 36–63% CD8+ T cells 35% 15–40% CD20+ B cells 31% 5–25% CD56+ NK cells 13% 6–30% T cell studies TRECs 1306 copies/0.5 mg DNA > 400 copies/0.5 mg DNA TCR v-beta Polyclonal (normal) Proliferation No mitogen 249 CPM 910 CPM PHA 6 mcg/mL 72,673 CPM 40,269 CPM PHA 25 mcg/mL 93,563 CPM 50,833 CPM CD3 mitogen 2059 CPM 25,200 CPM TREC, T cell receptor excision circle; CPM, counts per minute; PHA, phytohemagglutinin Italicized values reflect abnormal results J Clin Immunol (2019) 39:430–439 435

• Fig. 4 Peripheral blood immune alterations in CARMIL2-

immunodeficiency patient. CyTOF (n = 1) was completed on PBMCs from controls (n = 6), patients with UC (n = 2), and the patient with CARMIL2 mutation. a Representative viSNE plots of PBMCs obtained from the studied patients, followed by cumulative data (b) (NK-natural killer cells, Tregs-regulator T cells, MP-macrophages, nTh-naïve CD4 T cells, nB-naïve B cells, tB-transitional B cells, nTC-naïve CD8 T cells, E. M.Tc-effector memory CD8 T cells, EM-effector memory, DC-dendritic cells, DN-double negative, ILC-innate lymphoid cells)

a b c CD25 FOXP3

93.2% 52.4%

Control Patient

9.6% 0.2%

CD25 FOXP3

Control Patient

CFSE Max CFSE Max

• Fig. 3 Impaired T cell

proliferation and Treg generation in CARMIL2-immunodeficiency. a FACS plot (n = 1) showing decreased T cell proliferation of naïve CD4+ T cells of the CARMIL2-immunodeficiency patient (red) compared with healthy control (gray). In these studies, each peak going from the right to left corresponds to divi- sion of the cells, leading to dilu- tion of the CFSE. b In vitro Treg generation assay (n = 1) demon- strating diminished production of FOXP3+ Tcells from naïve Tcells

• f the patient, under polarizing
• conditions. c Treg frequency in

peripheral blood in the patient is reduced, compared with control, as measured by FACS (n = 1) 436 J Clin Immunol (2019) 39:430–439

eosinophilic esophagitis [4–7]. A susceptibility to EBV- related smooth muscle tumors and leiomyosarcoma was ob- served in some CARMIL2 patients [4, 5] and required com- prehensive follow-up. Although the patient we describe here presented with a psoriatic rash at an early age as consistent with the other re- ports, it resolved and did not recur. In addition, he did not suffer from significant respiratory illnesses and chronic viral infections, as often observed in patients with immunodeficien- cy disorders. Since colitis was the predominant feature of the boy’s ailment, he was regarded as an IBD patient and man- aged accordingly. Tissue eosinophilia in colonic biopsies is commonly seen in UC, and some studies have shown a negative correlation between the degree of eosinophilia and clinical outcomes [18, 19]. However, significant tissue eosinophilia is also a charac- teristic of several primary immunodeficiencies that are also associated with monogenic IBDs, such as IPEX syndrome (resulting from FOXP3 mutations) [20] and occasionally chronic granulomatous disease (resulting from CYBB, CYBA, NCF1, NCF2, and NCF4 mutations) [21]. Our patient, as well as those reported in the literature, exhibited severely diminished Tregs, similar to IPEX patients. Thus, CARMIL2- immunodeficiency should be added to the growing list of IPEX-like diseases, which, beside FOXP3, [20] also include mutations in IL2RA [22], MALT1 [23], STAT1 [24], CTLA4 [25], DOCK8 [26], and LRBA [27]. Interestingly, despite a lack of Tregs in our patient, no additional autoimmune features were noted. Tcell activation and subsequent signal transduction require co-stimulation of CD28. CARMIL2 was found essential for this co-stimulation process by linking CD28 with CARD11 (CARMA1) and activation of the NF-kB pathway [8, 28]. The human CARMIL2-immunodeficiency syndrome recapitulates the findings in mouse models, as demonstrated by the immu- nological workup in the patient described here and in previous

• studies. Studies in other patients showed global T cell immu-

nity impairment, with poor Th1 and Th17 differentiation and cytokine production, as well as abnormal activation of the NF- kB pathway [4, 6, 8]. While in mice the immune defects could be attributed to T cell dysfunction alone, in humans, the cel- lular defect appears to overlap both the T and B cells, leading to a combined immunodeficiency (CID) syndrome [6, 8]. Moreover, our CyTOF analysis provides a comprehensive

• verview of immune cell architecture, showing alterations in

the frequency of various other populations, including mono- cytes and NK cells. The IFN-γ expression was upregulated by different effector cells and might have an important role in mediating the inflammatory response. CARMIL2 contains five major domains (N- to C-terminal): pleckstrin homology (PH), LRR, helical dimerization (HD), capping protein-binding (CBR), and proline-rich domain

• Fig. 5 IFN-γ production in

CARMIL2-immunodeficiency. a Representative 2D CyTOF plot (n = 1) of IFN-γ production by effector memory T cells (CCR7−/ CD45RA−). b Heat map of IFN-γ production in the clusters gener- ated through automatic clustering

• f controls (n = 3), patients with

UC (n = 2), and patients with CARMIL2-immunodeficiency (n = 1) J Clin Immunol (2019) 39:430–439 437

(PRD) (Fig. 1c). The PH, LRR, and PRD have been previously demonstrated to play part in CD28 co-stimulation; therefore, impeding these regions leads to defects in T cell activation, as

• bserved in patients with CARMIL2 pathogenic variants, in-

cluding the patient described here [4–6, 8]. The LRR domain

• f CARMIL2 comprises 16 repeats of the canonical highly

conserved LxxLxLxxN/CxL sequence, in which BL^ corre- sponds with leucine, isoleucine, valine, or phenylalanine and BN/C^ with asparagine, threonine, serine, or cysteine [29]. The amino acid substitution p.Asn530Lys affects the highly con- served asparagine (N) in position 9 of the 11th LRR sequence (Fig. 1d); the alteration between the polar uncharged asparagine to a positively charged lysine, which also has a larger side- chain, at this position may possibly affect the α/β horseshoe/ loop structure of this domain, thus changing its affinity and interaction with other proteins, such as CARMA1 [8, 29]. The outcome from this amino acid substitution is decreased by 73% in protein expression in patient cells compared with a healthy control, highlighting its pathogenicity (Fig. 2a). These findings are similar to a different variant in the same LRR repeat (p.Leu525Gln), which was previously described in a patient of Turkish descent, without reported diarrhea, and shown to cause 65% decrease in protein expression [6]. Interestingly, immuno- histochemical analysis of control colonic and esophageal tis- sues shows expected high CARMIL2 expression in immune cells (lymphocytes and plasma cells), as well as high expression in endothelial cells and moderate expression in colonic epithe- lial tissue (Fig. 2c, Supplementary Fig. S1). These appear mark- edly reduced in the patient with mutated CARMIL2, i.e., no expression in colonic and esophageal immune cells and very mild expression in colonic epithelium and in endothelial cells in both tissues (Fig. 2d, Supplementary Fig. S1). These findings may explain the immunological gastrointestinal dysregulation and the clinical phenotype observed in this patient’s colon and

• esophagus. The observation of CARMIL2 expression in colon-

ic epithelial and endothelial cells warrants further study of the role of CARMIL2 in the gut. Overall, the phenotypical differences between the CARMIL2-immunodeficiency patients reported to date em- phasize the heterogeneity of clinical symptoms within the same genetic disease and occasionally even in patients with the same genetic variant. One of the X-linked variants ob- served in the patient’s WES, CYSLTR1 (NM_001282186.1): c.68A>G; p.Asn23Ser may have a modulating effect on his pronounced eosinophilic gastrointestinal disease. CYSLTR1 encodes a G protein–coupled receptor for cysteinyl leukotri- enes and has been previously associated with asthma and eo- sinophil migration [30]. However, the predicted effects of this variant are not strong (Table S1), and further evaluation is required to assess a possible effect on receptor activity. While the patient did have occasional eosinophilia on blood counts, his blood eosinophil levels are mostly normal and he does not have symptoms consistent with asthma. In summary, we describe a patient with infantile-onset co- litis that harbors a novel homozygous missense CARMIL2 variant, highlighting the importance of a wide differential di- agnosis in patients that present with presumably isolated coli-

• tis. The Treg signaling defects observed in this patient support

this diagnosis. Diagnosing CARMIL2-immunodeficiency in this patient has important clinical implications, requiring a multidisciplinary overview of gastroenterology and immunol-

• gy experts in patient management.

Acknowledgments We thank the patient and his family for participating in this study. Funding D.S.S. is supported by the Israel Science Foundation and Jeffrey Modell Foundation grants. S.B.S. is supported by NIH grants HL59561, DK034854, and AI50950; the Helmsley Charitable Trust; and the Wolpow Family Chair in IBD Treatment and Research.

Compliance with Ethical Standards

The study was approved by the institutional Helsinki committee, and written informed consent was obtained as customary. Conflict of Interest The authors declare that they have no conflict of interest.

References

1. Uhlig HH, Schwerd T, Koletzko S, Shah N, Kammermeier J, Elkadri A, et al. The diagnostic approach to monogenic very early

• nset inflammatory bowel disease. Gastroenterology. 2014;147:

990–1007. 2. Kammermeier J, Dziubak R, Pescarin M, Drury S, Godwin H, Reeve K, et al. Phenotypic and genotypic characterisation of in- flammatory bowel disease presenting before the age of two years. J Crohn's Colitis. 2017;11:60–9. 3. Uhlig HH, Schwerd T. From genes to mechanisms: the expanding spectrum of monogenic disorders associated with inflammatory bowel disease. Inflamm Bowel Dis. 2016;22:202–12. 4. Schober T, Magg T, Laschinger M, Rohlfs M, Linhares ND, Puchalka J, et al. A human immunodeficiency syndrome caused by mutations in CARMIL2. Nat Commun. 2017;8:14209. 5. Sorte HS, Osnes LT, Fevang B, Aukrust P, Erichsen HC, Backe PH, et al. A potential founder variant in CARMIL2/RLTPR in three Norwegian families with warts, molluscum contagiosum, and T- cell dysfunction. Mol Genet genomic Med. 2016;4:604–16. 6. Wang Y, Ma CS, Ling Y, Bousfiha A, Camcioglu Y, Jacquot S, et al. Dual Tcell– and B cell–intrinsic deficiency in humans with biallelic RLTPR mutations. J Exp Med. 2016;213:2413–35. 7. Alazami AM, Al-Helale M, Alhissi S, Al-Saud B, Alajlan H, Monies D, et al. Novel CARMIL2 mutations in patients with var- iable clinical dermatitis, infections, and combined immunodeficien-

• cy. Front Immunol. 2018;9:203.

8. Roncagalli R, Cucchetti M, Jarmuzynski N, Grégoire C, Bergot E, Audebert S, et al. The scaffolding function of the RLTPR protein explains its essential role for CD28 co-stimulation in mouse and human T cells. J Exp Med. 2016;213:2437–57. 9. Karczewski KJ, Weisburd B, Thomas B, Solomonson M, Ruderfer DM, Kavanagh D, et al. The ExAC browser: displaying reference 438 J Clin Immunol (2019) 39:430–439

data information from over 60 000 exomes. Nucleic Acids Res. 2017;45:D840–5. 10. Sherry ST, Ward MH, Kholodov M, Baker J, Phan L, Smigielski EM, et al. dbSNP: the NCBI database of genetic variation. Nucleic Acids Res. 2001;29:308–11. 11. The 1000 Genomes Project Consortium. An integrated map of genet- ic variation from 1,092 human genomes. Nature. 2012;491:56–65. 12. Scott EM, Halees A, Itan Y, Spencer EG, He Y, Abdellateef Azab M, et al. Characterization of greater middle eastern genetic variation for enhanced disease gene discovery. Nat Genet. 2016;48:1071–6. 13. Kumar P, Henikoff S, Ng PC. Predicting the effects of coding non- synonymous variants on protein function using the SIFTalgorithm. Nat Protoc. 2009;4:1073–82. 14. Schwarz JM, Cooper DN, Schuelke M, Seelow D. Mutationtaster2: mutation prediction for the deep-sequencing age. Nat Methods. 2014;11:361–2. 15. Adzhubei IA, Schmidt S, Peshkin L, Ramensky VE, Gerasimova A, Bork P, et al. A method and server for predicting damaging missense mutations. Nat Methods. 2010;7:248–9. 16. Ashkenazy H, Abadi S, Martz E, Chay O, Mayrose I, Pupko T, et al. ConSurf 2016: an improved methodology to estimate and visualize evolutionary conservation in macromolecules. Nucleic Acids Res. 2016;44:W344–50. 17. Biasini M, Bienert S, Waterhouse A, Arnold K, Studer G, Schmidt T, et al. SWISS-MODEL: modelling protein tertiary and quaternary structure using evolutionary information. Nucleic Acids Res. 2014;42:1–7. 18. Zezos P, Patsiaoura K, Nakos A, Mpoumponaris A, Vassiliadis T, Giouleme O, et al. Severe eosinophilic infiltration in colonic biop- sies predicts patients with ulcerative colitis not responding to med- ical therapy. Color Dis. 2014;16:O420–30. 19. Morgenstern S, Brook E, Rinawi F, Shamir R, Assa A. Tissue and peripheral eosinophilia as predictors for disease outcome in children with ulcerative colitis. Dig Liver Dis. 2017;49:170–4. 20. Bin Dhuban K, Piccirillo CA. The immunological and genetic basis

• f immune dysregulation, polyendocrinopathy, enteropathy, X-linked
• syndrome. Curr Opin Allergy Clin Immunol. 2015;15:525–32.

21. Posovszky C. Congenital intestinal diarrheal diseases: a diagnostic and therapeutic challenge. Best Pract Res Clin Gastroenterol. 2016;30:187–211. 22. Caudy AA, Reddy ST, Chatila T, Atkinson JP, Verbsky JW. CD25 deficiency causes an immune dysregulation, polyendocrinopathy, enteropathy, X-linked-like syndrome, and defective IL-10 expres- sion from CD4 lymphocytes. J Allergy Clin Immunol. 2007;119: 482–7. 23. Charbit-Henrion F, Jeverica AK, Bègue B, Markelj G, Parlato M, Avcin SL, et al. Deficiency in mucosa-associated lymphoid tissue lymphoma translocation 1: a novel cause of IPEX-like syndrome. J Pediatr Gastroenterol Nutr. 2017;64:378–84. 24. Uzel G, Sampaio EP, Lawrence MG, Hsu AP, Hackett M, Dorsey MJ, et al. Dominant gain-of-function STAT1 mutations in FOXP3 wild-type immune dysregulation-polyendocrinopathy-enteropathy- X-linked-like syndrome. J Allergy Clin Immunol. 2013;131:1611– 1623.e3. 25. Kuehn HS, Ouyang W, Lo B, Deenick EK, Niemela JE, Avery DT, et al. Immune dysregulation in human subjects with heterozygous germline mutations in CTLA4. Science. 2013;4:6–11. 26. Alroqi FJ, Charbonnier LM, Keles S, Ghandour F, Mouawad P, Sabouneh R, et al. DOCK8 deficiency presenting as an IPEX-like

• disorder. J Clin Immunol. 2017;37:811–9.

27. Charbonnier LM, Janssen E, Chou J, Ohsumi TK, Keles S, Hsu JT, et al. Regulatory T-cell deficiency and immune dysregulation, polyendocrinopathy, enteropathy, X-linked-like disorder caused by loss-of-function mutations in LRBA. J Allergy Clin Immunol. 2015;135:217–27. 28. Liang Y, Cucchetti M, Roncagalli R, Yokosuka T, Malzac A, Bertosio E, et al. The lymphoid lineage–specific actin-uncapping protein Rltpr is essential for costimulation via CD28 and the devel-

• pment of regulatory T cells. Nat Immunol. 2013;14:858–66.

29. Matsushima N, Tachi N, Kuroki Y, Enkhbayar P, Osaki M, Kamiya M, et al. Structural analysis of leucine-rich-repeat variants in pro- teins associated with human diseases. Cell Mol Life Sci. 2005;62: 2771–91. 30. Peters-Golden M, Henderson WR. Leukotrienes. N Engl J Med. 2007;357:1841–54. Publisher’s Note Springer Nature remains neutral with regard to juris- dictional claims in published maps and institutional affiliations. J Clin Immunol (2019) 39:430–439 439